Observed Signal · Sep 16, 2026 · Policy Update · Source: Astral Codex Ten · Impact: 2/5 · Sentiment: Positive
AI Superforecasting Could Transform FDA Drug Approvals
This guest post on Astral Codex Ten argues that AI superforecasting could transform FDA drug approval processes. The author, Josh Morrison of 1DaySooner, proposes a gradual implementation where the FDA first adds quantitative forecasts to existing decisions, such as predicting patient deaths or drug efficacy during trials. These forecasts would be checked against real-world data, eventually becoming public and then legal triggers for automatic approvals or trial progression. The post discusses the potential for reduced subjectivity, increased predictability for drug developers, and more efficient trials, while acknowledging the need for validation and calibration before full adoption.
This opinion piece discusses AI superforecasting for FDA reform, which is related to AI technologies but is not directly about AdTech/MarTech, and it's a proposal rather than a concrete development.
Track Real-Time AI & Technology Signals & Market Shifts
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Artificial intelligence superforecasters are described as becoming widely available within a few years.
- Josh Morrison, from advocacy group 1DaySooner, proposes using AI forecasting at the US FDA.
- The proposal includes a gradual three-stage implementation: adding forecasts to existing decisions, making forecasts public, and using them as legal triggers.
- The FDA's Sentinel program uses real-world evidence to identify safety signals among approved products.
- The Trump administration announced a pilot program to speed up phase 1 clinical trials under Operation Trialblazer.
- In 2021, the FDA faced criticism for its approval of Biogen's Alzheimer's drug Aduhelm.
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
AI Superforecasters Near Human Parity
The article examines the rapid rise of AI "superforecasters": specially scaffolded LLMs that research and produce probabilistic forecasts. Startups such as FutureSearch and Preseen claim strong returns on prediction markets and finance (e.g., FutureSearch produced a 7% forecast that US respiratory infections will halve by 2040; Preseen gave 8.8%). Metaculus benchmarking shows brand-name base models approaching crowd predictions as of May 2026, while forecasting-focused scaffolds add roughly nine months of effective model progress; scaffolded AIs are approaching top professional human forecasters and may reach parity within months. The piece discusses implications for finance, prediction markets, policy advice, and public trust, and argues that readily-available AI forecasters could democratize access to high-quality forecasting while reshaping markets where forecasting labor is valuable.
Agentic AI Transforming Pharma Marketing for Rare Diseases
This article discusses how agentic AI is poised to revolutionize pharmaceutical marketing, particularly for rare disease treatments. With traditional drug development costing $2.6 billion and taking 12-15 years, AI is compressing discovery timelines and enabling treatments for smaller patient populations, with rare disease approvals now exceeding 50% of FDA approvals. However, this precision medicine creates a marketing challenge: reaching small, specific audiences efficiently. The article argues that while programmatic advertising optimized campaign execution, agentic AI can synthesize vast healthcare, audience, and market data to provide actionable commercial insights, forecast patient need, and continuously adapt strategies. Early tests show marketing planning cycles collapsing by up to 10x. This approach aims to make more treatments commercially viable, improving treatment economics and driving further investment.
Amodei's Claim: AI Will Cure Most Diseases—Unrealistic
Gary Marcus (Substack) criticizes Anthropic CEO Dario Amodei's public claim that AI could "cure most human disease" within about five to ten years, arguing the timeline is naive given how medical science and drug development work. Marcus cites his earlier essay (and Emilia Javorsky's work) noting that, as of March, no AI-designed drug had reached clinical adoption. The piece collects pushback from multiple experts — including Eric Topol, Lior Pachter, Keith Robison, John Herrmann, Marios Georgakis, Ethan Mollick, and Dr. Catharine Young — who highlight the biological complexity of diseases (e.g., cancer's heterogeneity), the need for preclinical and clinical validation, and the current lack of causally sophisticated AI. Marcus concludes that Amodei's messaging risks giving false hope and recommends better calibration of medical claims.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.